JPH0750680A - Data transfer method for token passing type lan - Google Patents

Abstract

PURPOSE:To provide the dynamic allocation system of a data transfer band, which has an effect on resistance to fault and LAN efficiency. CONSTITUTION:A synchronous data quantity notice procedure S1 transmits synchronous data quantity notice frames 10 sharing measuring synchronous data quantity information of self nodes as token information or broadcast control information to other nodes, and similarly receives synchronous data quantity information from the other nodes. A synchronous data abolishment processing procedure S2 predicts the transmission lapse time of synchronous data concerned from the synchronous data quantity of the self and other nodes. Then, it immediately returns an approval response when the transmission request of a communication application can be terminated within different request time which can be designated to synchronous data concerned, and a refusal response when it cannot be terminated. A synchronous data transmission processing procedure S3 transmits synchronous data frames 20 which transmission-process and form synchronous data that is not abolishment-processed to the other nodes. Synchronous data concerned taking different request time which can be designated for synchronous data waiting for transmission from the communication application is abolished.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention is a token passing LAN (local area network) characterized by time criticality (timeliness of transferring data to a destination within a certain required time) required for factory automation and the like. Data transfer method.

[0002]

2. Description of the Related Art For example, as shown in FIG. 5, a conventional token passing LAN data transfer method disclosed in Japanese Patent Laid-Open No. 3-159436 discloses synchronous data (voice which requires severe time criticality in data transfer delay, etc.). Management node 2a (bandwidth management station) that manages bandwidth of high priority data) and destination node 2 as a destination of data transfer
b (calling station) and destination node 2 as destination of data transfer
Token-passing method LA that connects with c (the called station)
At N1, when the destination node 2b first receives a synchronous data transmission request from a communication application (communication application software realized by LAN), it transmits a band use request frame 11 requesting the transmission band of the synchronous data to the management node 2a. . Band use request frame 11 from the management node 2a
The request response frame 12 indicating reception of and the bandwidth use permission frame 13 indicating bandwidth allocation if possible (step S11 above). Next, when the destination node 2b receives the band use permission frame 13, the destination node 2b transmits a connection request frame 14 requesting the setting of a connection (logical connection) to the destination node 3c. If the acknowledgment frame 15 indicating the connection setting is received from the destination node 3c if possible, the synchronization data frame 20a is transmitted to the destination node 3c.
(Step S12 above). Destination node 2
When the transmission of the synchronous data frame 20a is completed, the b transmits a band return frame 16 indicating the end of band usage to the management node 2a. The return response frame 17 indicating the reception of the band return frame 16 is received from the management node 2a (the above procedure S13).

The token passing LAN is ISO
(International Organization for Standardization) Standard 8802 token ring type /
Token bus type and ISO standard 9314 token ring type FDDI (Fiber Distributed D)
In the data interface network, transmission right information called “token” circulates on the transmission path in one direction, and the node receiving the token controls the medium access to obtain the right to use the transmission path.

The conventional token passing system LA
The data transfer method of N refers to the management node every time the destination node receives the synchronous data so that the bandwidth utilization efficiency is not deteriorated even during the dynamic traffic (when the length of the synchronous data and the arrival interval are indefinite) The method of deciding allocation (data transfer bandwidth fixed allocation method) is adopted.

[0005]

In the conventional token passing LAN data transfer method as described above, the method of referring to the management node every time the destination node receives the synchronous data determines the transmission band allocation. , The complexity of the procedure for the countermeasure against the failure of the management node is increased and the scale is increased. Further, the overhead required for frame communication before and after the synchronous data transfer (procedures S11 and S13 in FIG. 5) increases and the responsiveness deteriorates. Further, the urgent synchronization data cannot be promptly transferred due to the token arrival waiting associated with the frame communication (step S11 in FIG. 5) before the transfer of the synchronization data. In addition, it requires a response time for determining the synchronous data transfer within the required time with reference to the management node every time, and the request and the response are described at different timings, which complicates the control program of the communication application. Furthermore, frame communication before and after synchronous data transfer (steps S11 and S13 in FIG. 5)
Complicates protocol (communication protocol) processing. In particular, it is necessary to add to the normal data transfer procedure a bandwidth usage status monitoring procedure such as unreturned allocated bandwidth due to node power failure and bandwidth return frame loss. The conventional data transfer method has the above problems.

[0006] The problem to be solved by the present invention is to eliminate the need for a management node in the token passing LAN data transfer method, measure the amount of synchronous data of its own node, and notify the other node to determine the amount of synchronous data of its own and other nodes. To provide a method (data transfer bandwidth dynamic allocation method) for deciding the discard and transmission processing of the synchronous data.

[0007]

In order to solve the above-mentioned problems, the token passing LAN data transfer method of the present invention is provided with the following procedure in the token passing LAN data transfer method for connecting a plurality of nodes:
It is characterized by adopting a data transfer bandwidth dynamic allocation method.

In the first procedure, each node transmits a synchronization data amount notification frame for measuring and storing the amount of synchronization data accumulated in its own node to another node at an appropriate timing.
The same synchronization data amount notification frame is received from another node. Also, the synchronization data amount information is piggybacked with the token information or the broadcast control information to form a concatenated or addition type token frame or a broadcast frame, and the notification transfer of the synchronization data amount information is performed.

The second procedure is that each node uses the synchronization data amount of its own and other nodes at an appropriate timing, and in response to the transmission request of the communication application, the estimated elapsed time until the synchronization data can be transmitted and the synchronization data LAN transfer. If the sum of the time delay time and the required time for the synchronous data exceeds, the synchronous data is discarded. If the transmission request of the communication application can be completed within the request time for the synchronous data, an affirmative response is returned, and if not, a rejection response is immediately returned. Further, a different request time can be specified for the synchronous data, which accompanies the transmission request of each synchronous data from the communication application.

In the third procedure, when each node arrives at the token, the synchronous data frame which is not discarded in the second procedure is transmitted and the synchronous data frame to be stored is transmitted to another node. In addition, the synchronization data for which the requested time has passed is discarded in the synchronization data waiting to be transmitted from the communication application. Further, the notification application can accompany each synchronous data transmission request and specify different request times for the synchronous data.

[0011]

According to the token passing LAN data transfer method of the present invention, each node first measures the synchronization data amount of its own node at an appropriate timing, and the synchronization data amount information is converted into token information or broadcast control information. Carry together to form a concatenated or token frame or broadcast frame, and send it to another node. The same amount of synchronous data amount information is received from other nodes (the above is the synchronous data amount notification procedure). Next, the transmission elapsed time of the synchronization data is predicted from the amount of synchronization data of the self and other nodes, and if the transmission request of the communication application can be completed within a different request time that can be specified for the synchronization data, an affirmative response, and if it cannot be rejected Immediately returns the response (above procedure for discarding synchronous data). When each node arrives at the token, the synchronization data transmission process that does not perform the discard process according to the discard process procedure is performed to form a synchronization data frame, and the synchronization data frame is transmitted to another node. Here, the synchronous data that has passed a different request time that can be specified for the synchronous data waiting to be transmitted from the communication application is discarded (the synchronous data transmission processing procedure).

[0012]

BEST MODE FOR CARRYING OUT THE INVENTION As shown in FIG. 1, a token passing LAN data transfer method according to an embodiment of the present invention is a token passing LAN 1 in which a plurality of nodes 2 are connected as data transfer destinations or destinations. (For example, A node 2), at an appropriate timing, the amount of synchronous data accumulated in its own node (a numerical value indicating the time required to transmit all the synchronous data, such as the number of synchronous data or the total number of octets, or a combination thereof). The synchronous data amount notification frame 10 for measuring and storing is transmitted to other nodes (for example, B and C nodes 2). The synchronization data amount notification frame 10 is received from the B and C nodes 2 in the same manner (the above is the synchronization data amount notification procedure S1). Next, A node 2
At an appropriate timing, the elapsed time until the synchronous data can be transmitted is predicted in response to the transmission request of the communication application using the amount of synchronous data of the self and other nodes. If the sum of the estimated elapsed time and the delay time at the time of transferring the synchronous data LAN exceeds the request time for the synchronous data, the synchronous data is discarded (the synchronous data discarding procedure S
2). Further, when the token arrives, the A node 2 transmits the synchronous data frame 20 that does not discard the synchronous data and stores the synchronous data frame 20 to the B and C nodes 2 (the synchronous data transmission processing procedure S3).

Token passing method LA of the above embodiment
The data transfer method of N is a method of measuring the amount of synchronous data of its own node, notifying the other node, and deciding the discard and transmission processing of the synchronous data from the amount of synchronous data of its own other node (data transfer bandwidth dynamic allocation method ).

As shown in FIG. 2, the synchronization data amount notification procedure S1 is a concatenated token frame (FIG. 2 (a)) or an addition type token frame (FIG. 2 (b)) in which the synchronization data amount information is combined with the token information. Broadcast frames (FIG. 2C) that are synergistic with the broadcast (broadcast) control information (header) are formed, and the notification transfer of the synchronous data amount information is performed between the nodes. Here, the token frame is a management frame essential for medium access control of the token passing LAN, whereas the broadcast frame is a data transfer frame destined to all nodes. When each node receives the token frame in its own node, in the case of the concatenated format of FIG. 2A, the sub-node corresponding to its own node in the concatenated subfields of the number of synchronization data in the first to Nth nodes (N is the maximum number of nodes) Write a lot of synchronization data in your node in the field,
It is transmitted to other nodes as the synchronization data amount information. On the other hand, the number of synchronization data in another node is read from the subfield corresponding to another node and stored in the storage area corresponding to the node (the number of synchronization data in the own node can be known from the number of synchronization data waits in the own node). In the case of the addition type format shown in FIG. 2B, the numerical value read from the synchronous data number cumulative field and the synchronous data number in the own node are added, and written in the synchronous data number cumulative field, and the other node as synchronous data amount information. Send. On the other hand, the difference between the numerical value read from the relevant synchronous data number cumulative field before the current node writes this time and the numerical value previously written to the relevant synchronous data number cumulative field by the local node as the other node's cumulative synchronous data number Store in the storage area for the number of synchronous data. In addition, each node uses the node number of its own node and the number of synchronization data waiting to be transmitted, as in the case of the broadcast frame format in FIG. 2C, at each appropriate transmission timing (for example, when the token circulation is performed an appropriate number of times before token release). Is stored in each field and is transmitted to other nodes as synchronization data amount information. On the other hand, when a broadcast frame is received from another node, the number of synchronous data waiting to be read, which is read from the field, is identified by the node number field and stored in the storage area corresponding to the node. It should be noted that instead of the number of synchronous data waiting to be transmitted, another amount of synchronous data such as the total number of octets of synchronous data may be used.

In the synchronous data discarding procedure S2, as shown in FIG. 3, each node first receives a synchronous data transmission request from the communication application (procedure S101). Next, the cumulative total of the number of synchronous data waiting for transmission in the inode (Qi) (ΣiQi, i
= 1 to N) and the average delay time (d) when transferring the synchronous data LAN and the required time (T) for the synchronous data.
Are compared with each other to determine whether the transmission of the synchronous data can be completed within T (step S102). Further, if it can be completed within T, the synchronous data is added to the synchronous data queue (queue), and a positive response is returned to the transmission request of the communication application. If you can't finish within T,
The synchronous data is discarded, and a rejection response is returned (above procedure S
103, S104).

In the synchronous data transmission processing procedure S3, as shown in FIG. 4, each node first receives a token frame from the LAN, and determines whether or not there is synchronous data waiting for transmission in the own node (the above procedure S201, S202). . Next, if there is transmission-waiting synchronization data in step S202, the current time (t) and the average delay time (d) during the synchronization data LAN transfer
The time (a) passed from the communication application to the synchronous data is subtracted from the sum of the difference and the difference between the difference and the required time (T) for the synchronous data, and the transmission of the synchronous data ends within T. Determine if you can. When it can be completed within T, the synchronization data is transmitted to the LAN. When the processing cannot be completed within T, the synchronization data is discarded and the communication application is notified of the discard (the above steps S203 to S203).
205). Further, the presence / absence of synchronous data in the transmission waiting state in its own node is determined (step S206). If not, the asynchronous data transmission process is performed and the token is released (steps S208, S209). If there is, a comparison is made between the time elapsed after the node has received the token frame and the token holding time (defined by the token passing LAN protocol) to judge. If the elapsed time exceeds the token holding time, the token is released. If not,
The determination processing procedure of step S203 is repeated (above step S2
07, S209). On the other hand, if there is no transmission waiting synchronous data in step S202, asynchronous data in a transmission waiting state within the own node (low priority data such as file transfer that does not require time criticality that is not severe in data transfer delay)
And the token is released (step S208,
S209).

Synchronous data discard and transmission processing procedures S2 and S3
Then, the communication application passes the request time (T) for the synchronous data to each node along with the transmission request for the synchronous data. Even if a different request time is specified for each synchronization data, each node holds the request time and the request time passed from the communication application, sends the synchronization data at the appropriate timing within the request time, and discards it if not possible. To do.

[0018]

According to the token passing LAN data transfer method of the present invention as described above, the amount of synchronous data of its own node is measured and notified to another node, and the synchronous data is discarded from the amount of synchronous data of its own other node. Therefore, compared to the conventional method of deciding the transmission band allocation by referring to the management node every time the source node receives the synchronous data, it is more fault-tolerant and can be easily implemented.
Further, the utilization efficiency of the LAN is improved and the delay in transferring the synchronous data LAN can be reduced. In addition, the synchronization data amount information is piggybacked on the token information and the broadcast control information to minimize the use of the LAN band associated with the synchronization data amount notification.
It can be easily designed with a commercially available LAN control LSI compliant with O8802 or 9314 standard LAN. Further, it facilitates the control program of the communication application, such that it can immediately respond to the synchronous data transmission request of the communication application and can be described synchronously. Furthermore, there is an effect that the communication application can specify a different request time for each synchronous data to flexibly configure the system on the LAN.

[Brief description of drawings]

FIG. 1 is a sequence diagram of a token passing LAN data transfer method according to an embodiment of the present invention.

FIG. 2 is a format diagram showing a concatenated / additional token frame and a broadcast frame in which synchronous data amount information is combined with token information and broadcast control information.

FIG. 3 is a flowchart showing a procedure for discarding synchronous data shown in FIG.

FIG. 4 is a flowchart showing a procedure of a synchronous data transmission process shown in FIG.

FIG. 5 is a sequence diagram of a conventional token passing LAN data transfer method.

[Explanation of symbols]

 1 LAN 2 Node 10 Synchronous data amount notification frame 20 Synchronous data frame In the drawings, the same reference numerals indicate the same or corresponding parts.

 ─────────────────────────────────────────────────── ─── Continuation of the front page (72) Inventor Kazuki Odaka 5-1-1, Ofuna, Kamakura City Mitsubishi Electric Corp. Communication Systems Research Laboratories (72) Inventor Yuji Atsui 5-1-1, Ofuna, Kamakura City Mitsubishi Electric (72) Inventor Tetsuo Ideguchi 2-3-3 Marunouchi 2-3, Marunouchi, Chiyoda-ku, Tokyo Sanryo Electric Co., Ltd.

Claims (6)

[Claims] 1. In a token passing LAN data transfer method in which a plurality of nodes are combined, a synchronization data amount notification frame for measuring and storing the amount of synchronization data accumulated in each node by each node at an appropriate timing is transmitted to another node. To the communication request from the other node using the first procedure for receiving the synchronization data amount notification frame from the other node in the same manner and each node using the synchronization data amount of its own and other nodes at appropriate timing. A second procedure for discarding the synchronous data if the sum of the estimated elapsed time until the synchronous data can be transmitted and the delay time at the time of transferring the synchronous data LAN exceeds the request time for the synchronous data; When the token arrives, the synchronization data frame that does not discard the synchronization data is sent and the stored synchronization data frame is sent to another node. And a third procedure for performing the token passing LAN data transfer method. 2. The synchronization data amount information is piggybacked on the token information in the first procedure to form a concatenated or addition type token frame, and the notification transfer of the synchronization data amount information is performed. Token passing LAN
Data transfer method. 3. The token passing method LA according to claim 1, wherein in the first procedure, the synchronous data amount information is piggybacked with the broadcast control information to form a broadcast frame, and the notification transfer of the synchronous data amount information is carried out.
N data transfer method. 4. The method according to claim 1, wherein an affirmative response is immediately returned if the transmission request of the communication application can be completed within the request time for the synchronous data in the second procedure, and a rejection response is immediately returned if the transmission request cannot be completed. Token passing method LAN data transfer method. 5. The data transfer method for a token passing LAN according to claim 1, wherein in the third procedure, among the synchronization data waiting to be transmitted from the communication application, the synchronization data for which a required time has elapsed is discarded. . 6. The token passing system according to claim 1, wherein the communication application accompanies the transmission request of each synchronous data by the second and third procedures, and different request times can be designated for the synchronous data. LAN data transfer method.